VBR7 Strategy - Comments needed

[kellogs]

Dear Members,

I have a VBR7 which is installed in a physical server. This VBR7 server is on the same local network as my the other 3 ESXI 5.1U1 hosts. 2 Of the ESXI5.1U1 hosts is setup as a 2 nodes VSA and the 3rd ESXI 5.1U1 hosts is added to the VSA Cluster and mounted to both VSA Datastores and it has its own local data store as well.

I have an offsite NAS which is on a 500Mbps link to this DC.

I am planning to

1. replicate all VMs from these 2 VSA ESXI 5.1u1 hosts to the standalone ESXI5.1 host local data store
2. backup all the VMs to a mounted external NFS server (same data center) - 10G link
3. backup all the VMs to a mounted remove NAS - NFS server - 500Mbps link

There are about 50 VMs with a mixture of Windows and Linux and storage used size is about 15TB

May i know if this was your baby, what would be your strategy on the backup and replicate jobs?
How many replication retention point would you keep?

Btw we already have a file level based backup to restore file/folders on the fly.

Thanks much!

[foggy]

May i know if this was your baby, what would be your strategy on the backup and replicate jobs?
How many replication retention point would you keep?

Your setup looks good to me, however the particular job settings entirely depend on RPO and RTO values accepted/required in your company. You can also search this forum for some general recommendations, like grouping VMs having similar OS'es in one job for better deduplication.

I would also suggest using backup copy jobs to send backups to the remote site to avoid snapshotting production VMs twice.

[kellogs]

Dear Foggy,

Do you recommend that I would group replication job based on the VM OS?

Linux VMs in its own Replication Job and Windows VMs in its own Replication Job?

Do you recommend to run both scheduled jobs at the same time?

I noticed that Veeam only runs 2 tasks parallel per job. Is this normal?

[veremin]

Do you recommend that I would group replication job based on the VM OS?

For backup jobs, we typically recommend grouping VMs with the same OS within one job. As mentioned, this should guarantee better deduplication ratio. Since replication restore points are stored in uncompressed/undeduped state, you can group VMs however you like to.

Do you recommend to run both scheduled jobs at the same time?

If your backup infrastructure can cope with the load caused by replication activity, then, you can run them simultaneously.

I noticed that Veeam only runs 2 tasks parallel per job. Is this normal?

What about your proxy setup? How many proxy servers you have? What specification they have? Do proxy servers involved in backup/replication job have any predefined number of concurrent tasks?

Thanks.